The urgency of the problem of short-term forecasting of electricity imbalances in the conditions of the modern electricity market of Ukraine is substantiated. A comparison of the results of forecasting daily graphs of electricity imbalances using autoregressive models ARIMA, VARMA and developed on their basis combined models with the influence of predicted values of generation of renewable sources. Analysis of the obtained results shows that the VARMA vector autoregression model has accurate results. References 11, figures 2, tables 2.